Hetare King

I’m not sure that the comparison with the weather data works. Tweaking curves to more closely match the test data, and moving around a model’s probability space in the hope that it sufficiently increases the probability of outputting tokens that fixes the code’s problems, seem different enough to me that I don’t know whether the former working well says anything about how well the latter works.

If I understand what you’re describing correctly, the two models aren’t improving each other, like in adversarial learning, but the adversarial model is trying to get the generative model to zone in on output that produces the user’s desired behaviour based on the given test data. But that can only work as well as how much the adversarial model can be relied upon to actually perform the tasks needed to make this happen. So I think my point still stands that the objectivity of your measurements of the test results is only meaningful if the test results themselves are meaningful, which is not guaranteed given what’s doing the testing.

How complex is the adversarial model? If it’s anywhere near the generative model, I don’t think you can have actual meaningful numbers about its reliability that allow you to reason about how meaningful the test results it produces are.

Sorry, that still doesn’t really make sense to me. If you can’t trust the generative model to produce code that does what it’s supposed to do, then you also can’t trust the adversarial model to perform the tests needed to determine that the code does what it’s supposed to do. So if the results have no meaning, then the fact that you can objectively measure them also has no meaning.

Programmers are kind of weird in the context of this post, because we tend to pretty consistently think our job is simpler than it really is, despite constantly being proven wrong.

objectively testable by adversarial models

This is an odd thing to say. Adversarial models are still learning models and have all the limitations that implies, including objectivity being far from guaranteed.

That sounds more like tinkering around the edges to me. Whipping companies like Twitter into behaving, while it absolutely needs to happen, won’t fundamentally change anything about the dependency of Europe to those companies and the pressure the US can exert through that dependency.

It’s an inevitable outcome of its structure. With memes, it’s usually just the low-information image, which is typically visible from the post listing. There’s no article to read, no video to watch (or just a very short one), no question to think about, and you can upvote it straight from the post listing, so there’s not even a link to click. In other words, memes have a very low barrier-to-upvote compared to other types of posts, and as a result, are more likely to get upvotes and end up on the front page.

For serious conversation what you really want is a forum or only join communities on Lemmy where memes are frowned upon.

• Princess Tutu
• Legend of the Galactic Heroes (the original, though the newer Die Neue These is decent too)
• Mahoujin Guru Guru (1994 and 2017)
• King of Braves GaoGaiGar
• Slayers

“But the plans were on display…” “On display? I eventually had to go down to the cellar to find them.” “That’s the display department.” “With a flashlight.” “Ah, well, the lights had probably gone.” “So had the stairs.” “But look, you found the notice, didn’t you?” “Yes,” said Arthur, “yes I did. It was on display in the bottom of a locked filing cabinet stuck in a disused lavatory with a sign on the door saying ‘Beware of the Leopard.”

Despite the quality of their results going down in recent years and getting worse because of AI slop, the search engines I would miss the most in terms of type of service. Most alternative search engine still use the indices of Google and/or Bing and the ones that don’t, don’t have a very big index. I’m old enough to remember a time when search engines were plentiful, but terrible, and back then I actually made use of web directories, like Yahoo! at the time, more. A still-existant example would be Curlie, an heir to dmoz, and there are also more local sites like the Dutch Startpagina. Being more dependent on things like that would probably make my web usage more exploratory and less about trying to find a specific piece of information quickly. And I would also go directly to specific websites more often when I do need specific information. But there are also a few companies working on making a European search index and this happening would undoubted accelerate their efforts, so depending on how that works out, not much might change at all.

Streaming-wise, there are local streaming services for films and TV shows and they would undoubtedly expand their offerings with the loss of competition from American giants, but also, I never stopped buying BDs and DVDs (in fact I have a backlog). I never understood the appeal of music streaming, so I still buy music, sometimes even on CD. As for something like YouTube, Nebula is America-based, but it’s not “big tech”, so I would watch more of that. Niconico Douga isn’t what it used to be, but that might change without YouTube. And there would probably also be some movement towards federated video streaming.

I don’t actually make use of any of the big social media platforms. Technically, I have a LinkedIn account, but I don’t really use it and wouldn’t miss it. It’s not really social media, but I do use WhatsApp, but that being gone would just make it easier to convince friends and family to switch to something better.

It does have a USB port, but it’s better not to connect any weird devices to your PC. But if you have an old laptop or Raspberry Pi or something that you’re willing to potentially sacrifice (making sure it’s free of any personal data and not connected to any network), you could maybe find out what it’s describing itself as.

I imagine his thinking went something like this: “If they build cycling infrastructure, it would attract casual cyclists. But I’d be zooming past them at several times their speed, which would be considered dangerous, so speed limits would be implemented and I wouldn’t be able to use public infrastructure as a practice track anymore.”

I’m not sure I entirely buy that. For cloud gaming to be any good at all, you need a high-speed, low-latency internet connection. Yes, nowadays having an internet connection is pretty much a requirement in the industrialised world and even someone of lesser means will probably have one good enough to watch streaming video at a decent enough quality (unless they live in the middle of nowhere), but that’s not good enough. So with the expensive internet connection and the monthly subscription, cloud gaming doesn’t strike me as a very economical.

We’ve also been living in a period of diminishing returns when it comes to visual fidelity improving as hardware power does for a while now, so you can buy older, more affordable hardware and still have games look great on them. Meanwhile, I don’t think someone who insists on being able to see the surroundings accurately reflected in every window and puddle is going to accept the compression artifacts and latency of cloud gaming.

PNG does not compress photos very well. A photo that is 5 MB when saved as a high-quality JPEG may very well be at least 15 MB as a PNG. Also, a lot of cameras (phone or otherwise) save to JPEG by default.

I do wish more people would use PNG where it makes sense, though. The other day I made an edit to an image containing line art that was purely black and white except for the compression artifacts. I applied a threshold so that all the artifacts became either perfectly black or white and saved it as a monochrome PNG, reducing the file size to less than a third, while containing more information and having a cleaner image. I later remembered that I could reduce the file size even more by using indexed colours. In other words, whoever originally saved it in a lossy format actually made it take up more space than needed while also needlessly reducing image quality.